Classification with Mixture of Experts Models

Loading...
Thumbnail Image

Persistent link to this item

Statistics
View Statistics

Journal Title

Journal ISSN

Volume Title

Title

Classification with Mixture of Experts Models

Published Date

2022-12

Publisher

Type

Thesis or Dissertation

Abstract

Mixture of experts (MoE) layers allow for an increase in model parameters without acorresponding increase in computational cost by utilizing sparse dynamic computation across “expert” modules during both inference and training. In this work we study whether these sparse activations of expert modules are semantically meaningful in classification tasks; in particular, we investigate whether experts develop specializations that reveal semantic relationships among classes. This work replaces the classification head of selected deep networks on classification tasks with an MoE layer. MoE layers allow for the experts to specialize in ways that are qualitatively intuitive, and quantitatively match structural descriptions of their relationships better than the classification heads in the original networks.

Keywords

Description

University of Minnesota M.S. thesis. December 2022. Major: Computer Science. Advisor: Dongyeop Kang. 1 computer file (PDF); vi, 35 pages.

Related to

Replaces

License

Series/Report Number

Funding information

Isbn identifier

Doi identifier

Previously Published Citation

Suggested citation

Mooney, James. (2022). Classification with Mixture of Experts Models. Retrieved from the University Digital Conservancy, https://hdl.handle.net/11299/252470.

Content distributed via the University Digital Conservancy may be subject to additional license and use restrictions applied by the depositor. By using these files, users agree to the Terms of Use. Materials in the UDC may contain content that is disturbing and/or harmful. For more information, please see our statement on harmful content in digital repositories.